These slides are adapted from a presentation written by Bryan C. Hassel and Emily AyscueHassel and delivered by Bryan C. Hassel at a meeting held by the Center for AmericanProgress on defining and measuring teacher effectiveness.

1

We care about �“teacher effectiveness�” because we care about each teacher�’s �“effect�” onstudents.

So today what I�’m proposing is actually a simple formula for thinking about a teacher�’s�“effect.�”

A teacher�’s effect equals the teacher�’s effectiveNESS times the teacher�’s REACH.

EffectiveNESS �– how MUCH do the teacher�’s students learn, on average �–

Times REACH �– how MANY students receive instruction from the teacher.

�“Reach�” is an essential part of the formula because a highly effective teacher who teachesa lot of students has a greater effect than one who teaches just a few.

I�’m going to talk first about effectiveness, then about reach.

2

The core of a state�’s definition of teacher effectiveness must be student outcomes:

How much students learn, and other valued outcomes.

And that goes for every subject and grade, whether or not it�’s currently assessed with anofficial standardized test. EVERY teacher�’s effectiveness ultimately hinges on her students�’outcomes.

I�’m not going to talk about how to MEASURE student outcomes that�’s the topic ofsubsequent sessions. My point here is just that student outcomes need to be the core,because no other core will help teachers and schools meet the ambitious goals we all sharefor improving student outcomes.

By saying �“core�” I don�’t mean that outcomes are ALL that needs to go into a definition ofteacher effectiveness. The definition also needs to include teacher behaviors. Outcomesare still core, because it�’s behaviors LINKED TO OUTCOMES that need to be part of thedefinition.

But why include behaviors, if student outcomes are paramount?

3

Because outcome measures don�’t tell us everything we need to know.

First, they don�’t yet exist for many grades, subjects and other valued outcomes.

Second, they don�’t yield development advice for teachers. If a teacher falls short of goals,the outcome measures don�’t tell the teacher or others WHY he fell short or what he needsto do differently in the future

Third, they don�’t tell us who should specialize or take on advanced roles. A teacher mayscore very high on outcome measures, but that doesn�’t mean she will be a great leader ofother teachers, curriculum developer, or any other advanced role we might envision.

And similarly, outcome measures can�’t be used to select new teachers �– who obviouslyhaven�’t yet had the chance to produce student outcomes.

4

So, states need to include teacher behaviors in their definitions of teacher effectiveness.

When including behaviors, though, states need to do it well. Here are four essentialguidelines that every state should keep in mind.

One �– choose wisely, selecting only those behaviors that are correlated with valuedoutcomes. We need to identify actual teachers who achieve the highest outcomes, andthen use rigorous research to identify the behaviors that set them apart from average andlow performers.

I know that sounds difficult, but this is the only valid way to select behaviors for inclusion ina definition of teacher effectiveness. Expert opinion, focus groups, consensus processes allsound good �– but they are poor substitutes for analysis of actual behaviors linked tooutcomes. That�’s the standard every state should apply.

5

Second �– once states have identified behaviors to include, they need rigorousmeasurement to assess how well teachers are engaging in behaviors that matter. Again,I�’m not getting into measurement. My point here is just that there�’s no point in deployingelaborate systems to measure teacher behavior without also investing in great tools andprocesses that make those measures reliable.

Third is dynamic improvement. Identifying behaviors that matter, and measuring themreliably, can�’t be a one time project. We shouldn�’t wait until we have perfect measures tostart this work. But then both identification and measurement need to be continuouslyimproved, not etched in stone in the spring 2010 legislative session as everyone�’s preparingfor Race to the Top.

Fourth, states need to look beneath the surface, seeking deeper competencies thatdifferentiate performance even among teachers who seem similar on the surface. [go rightto next slide]

6

Think of a teacher�’s capability as an iceberg, with a tip above the water line that�’s easilyseen �– but a giant mass below the water that�’s not as easily observed.

The very tip is the teacher�’s formal credentials, like certification and degrees. These arethe easiest to see, but we know from years of research that they explain very little of thedifferences in outcomes that teachers produce.

Also above the water line are a teacher�’s knowledge and skills, and her easily observedpractices.

Below the water line are deeper competencies that aren�’t as obvious based on shortobservations of teachers in action �– the recurring, underlying patterns of thought andaction that underlie more easily observed practices. It�’s easiest to see this with anexample (see next slide).

7

Start with Teacher A. At the beginning of the year, he�’s eager to do well. He�’s attentive tothe state curriculum and the district�’s pacing guides. He follows the teaching practices laidout in the state�’s new definition of teacher effectiveness. The district requires benchmarkassessments every 2 months, and he dutifully administers them and reviews the data.When a student has trouble, he�’s concerned �– and when he has time uses his knowledge tohelp the student as well as he can.

But now think of a second teacher. At the beginning of the year, she spends extra timeconsidering where each of her incoming students is starting. She pores over all of lastyear�’s data, and checks in with the students�’ prior teachers to learn more about thebarriers each child faces. Based on that, she does research and talks to peers to prepare totackle those specific challenges. She�’s kept track of past experiences with similar kids, andshe has that to go on, too. She assesses her students every week, not just the mandatedevery two months. Every week she knows who�’s making progress and who�’s stuck. Basedon her knowledge and new research she does on new barriers, she makes a plan to try newtactics �– and executes it. And when there�’s a student�’s need she can�’t meet, she gets helpearly and often, from her colleagues or from outside the school. She doesn�’t give up.

If you observed these two teachers for half an hour, you might pick up on these differences.But what really distinguishes these 2 �– advance planning, hard analysis & thinking, learningover time, and relentless adjustment and follow up, are much harder to observe in shortstretches.

8

That doesn�’t mean they CAN�’T be observed. In fact there are well honed tools foridentifying, measuring and using these deeper competencies in a variety of professionalsettings. So there�’s no reason why education should be looking just at the tip of theiceberg of teacher capabilities when there�’s such an enormous mass of deepercompetencies beneath the surface that in all likelihood better explain performance.

9

Now �– whether we�’re talking about student outcomes produced by teachers, or teacherbehaviors, we�’re likely to observe a performance curve like this, just like we would see inany job.

Most teachers are in the Great Middle. And one of the big reasons our debates aboutdefining and measuring teacher effectiveness get so heated is because of worries thatmeasures within this great middle won�’t be accurate and fair. That�’s because it actually istechnically difficult to tell a definitive difference between teachers who are, say, exactlyaverage and those that are a few points above average.

The two ends, or tails, of the curve are another matter. The lowest and highest performersstand out from the crowd. And, their presence in schools makes an enormous differencefor kids �– the best are way better than the norm, and the worst teachers are way worse.

10

These two qualities of the �“tails�” make them the best focus for our efforts to makedistinctions in teacher effectiveness. It�’s technically more feasible to identify teachers inthe high and low ends than to make fine distinctions within the Great Middle.

AND the learning payoff is also much higher of knowing who those highest and lowestperformers are.

So what is that payoff? As Tom Kane and his colleagues have shown, removing a higherproportion of the lowest performers could have a huge effect on student outcomes.

Identifying very high performers could have a similar payoff �– if we could retain more ofthe very best teachers for longer, we�’d see large benefits for kids.

But there�’s an added potential benefit of identifying the very best teachers �– and that�’sextending their reach to more students �– which brings me back to�….

11

My original simple formula �– states need to know not just whether an individual teacher ishighly effective, but how many students are being reached by that highly effective teacher.

12

That�’s because if states know who their best teachers are, and their current reach, they canfocus on �“reach extension�” which is just increasing the number of students who receive topinstruction.

Now, there are many ways schools, districts, and states could extend the reach of theirbest. They can do it by using technology to give students access to great teachers nomatter where they live.

Or they can do it by reorganizing teachers�’ time and roles in schools so that the bestteachers interact with more students.

For more on reach extension, see Public Impact�’s working paper 3X for All:Extending the Reach of Education�’s Best, October 2009 available at:http://publicimpact.com/publications/3xforall_extendingthereachofeducationsbest_publicimpact_october2009.pdf

13

Ultimately reach extension is so important because it�’s a largely untapped way of gettingthe best instruction to more kids.

Consider 2 districts, both with 50,000 elementary students, and both with 500 teachers inthe highest effectiveness category of your state�’s new teacher effectiveness system.

But what if in district A, these top teachers only reach 18 students each, while in district Bthey reach 27.

Now let me pause because you may be thinking I�’m talking here about increasing class sizesfrom 18 to 27, but I�’m not necessarily talking about that. Let me give you an example.Right now, elementary teachers spend about half their time on non instructional duties.What if, once we know a teacher is in the top effectiveness category, she�’s relieved of justabout all of those non instructional duties. Then, she could teach 50% more kids and stillhave time left over to coach her peers, or follow up with individual students needing morehelp.

So in this example, District B reaches 50% more kids with top instruction �– without havingany more highly effective teachers. And then imagine these numbers on a national scale.And then consider another payoff of reach extension �– as more top teachers enter and stayin the profession because of the opportunities for career advancement and additional paythat reach extension enables.

14

Now there are a lot of big implications for states in this discussion of defining andmeasuring teacher effectiveness, and we�’re going to spend some time as a group later inthe day really delving into that.

But here let me suggest three of the big roles states need to assume.

They first need to create the will to focus on teacher effectiveness. They have a couple ofprimary tools for doing this. One is requiring teacher effectiveness measures that trulydifferentiate teachers based on their performance, with student outcomes as the core.And then second, shining a bright light on a range of metrics that reveal how differentschools and districts are doing on improving teacher effectiveness.

Second, states need to lead the way on the whole range of activities we�’re talking abouthere �– from identifying the student outcomes and teacher behaviors that matter, todevising the measures of these elements, to ensuring the use of the resulting data acrossthe whole range of teacher policies. Some districts and schools will want to go above astate mandated floor on these points, but the state�’s role in creating that floor is critical.

Finally, states need to stay the course to improve identification of what matters,measurement, and the use of teacher effectiveness data. As I said before, this can�’t be aone time project.

15

That�’s especially important because as soon as all states move from the current abysmal1.0 systems of teacher evaluation to version 2.0, the need for version 3.0 is going to be onthe horizon.

Teachers today largely work in a one teacher, one classroom world. Even the very best havefew career opportunities that keep them teaching. Technology is an add on.

But there�’s some indication that all of this is about to change. And if it does, the 2.0systems we�’re all working on right now are going to need to change too. So even if everystate �“races to the top�” next year in defining and measuring teacher effectiveness, we allneed to realize �– we�’re not done.

16

17